有时候不用的指标的绝对值不能比，但是转转为百分比的形式就容易看出波动了，是数据分析的好用的一个分析函数
20:00:24 sys@orcl> conn scott/tiger;
connected.
20:00:30 scott@orcl> create table test
20:01:22 2 (
20:01:22 3 name varchar(20),
20:01:22 4 kemu varchar(20),
20:01:22 5 score number
20:01:22 6 );
table created.
elapsed: 00:00:00.04
20:01:23 scott@orcl> insert into test values('testa','yuwen',10);
1 row created.
elapsed: 00:00:00.02
20:01:35 scott@orcl> insert into test values('testa','英语',100);
1 row created.
elapsed: 00:00:00.00
20:01:35 scott@orcl> insert into test values('testb','yuwen',60);
1 row created.
elapsed: 00:00:00.01
20:01:36 scott@orcl> insert into test values('testb','yuwen',120);
1 row created.
elapsed: 00:00:00.00
20:01:36 scott@orcl> insert into test values('testc','yuwen',40);
1 row created.
elapsed: 00:00:00.00
20:01:37 scott@orcl> commit;
commit complete.
elapsed: 00:00:00.00
20:01:42 scott@orcl> select name,
20:03:32 2 score,
20:03:32 3 ratio_to_report(score) over() as ratio1,
20:03:32 4 ratio_to_report(score) over(partition by kemu) as ratio2
20:03:32 5 from test ;
name score ratio1 ratio2
-------------------- ---------- ---------- ----------
testa 10 .03030303 .043478261
testb 60 .181818182 .260869565
testc 40 .121212121 .173913043
testb 120 .363636364 .52173913
testa 100 .303030303 1
elapsed: 00:00:00.00
20:03:33 scott@orcl>
总结
以上所述是www.887551.com给大家介绍的oracle百分比分析函数ratio_to_report() over(),希望对大家有所帮助